One Size Doesn’t Fit All: The Limitations of Assuming Data Consistency in Health Spending

The Health Affairs Blog today published a piece from RTI International, RTI Health Solutions and National Pharmaceutical Council researchers detailing the assumptions made – and the blind spots created – by a recent Journal of the American Medical Association piece by Papanicolas et al. suggesting that the United States spends more on health care than other countries because of higher prices, higher health care salaries, and administrative costs.

A more detailed look at the data underlying the claim, from the Organisation for Economic Co-operation and Development (OECD), underscores the benefits and challenges of the dataset and making cross-country comparisons. These challenges fall into four categories:

  1. OECD data does not consider the intensity of services and lumps all U.S. health utilization together, which obscures a critical reality: the enormous variability in the U.S. system means there is no meaningful “average” when it comes to health care use.
  2. The OECD data provides a snapshot of two dozen frequently used services, far less comprehensive than all services used to treat all health conditions.
  3. These data do not include prices for any service category and infer prices based on expenditures. While efforts were made to overcome this issue in the JAMA article, publicly available prices for services typically reflect charges rather than actual price paid.
  4. Similar to the health care use, the price paid varies, at a minimum, by geographic region, payer, health system, and provider.

One of the article’s authors, Leslie Greenwald, PhD, MPA, chief scientist and vice president, RTI International, explains why we need to dig deeper when examining these data. There is a great deal of value in analyses such as those performed for the Papanicolas paper. But the limitations of the underlying data and the idiosyncrasies of the world’s largest health care market should be considered when drawing conclusions and developing policy proposals based on results from conjecture or inference rather than from actual data.


This content was originally published as part of Going Below The Surface, an NPC-led initiative to broaden and improve the conversation around how we use health care resources in the United States.